Snowflake, a cloud data warehouse vendor, recently announced Data Sharing, which enables any Snowflake customer to share any part of their data warehouse with other Snowflake customers. Compared to traditional data sharing techniques, Snowflake’s method is simple because no data is moved or copied.

Traditional Methods. Typically, organizations that share data, whether internally with different divisions or externally with customers or partners, can either use replication, file transfers, application programming interfaces, or email. However, these approaches are suboptimal since in all cases the data must be duplicated, physically moved, and stored in a separate database. Plus, keeping data secure and synchronized is difficult.  

Snowflake’s Method. Snowflake’s new Data Sharing feature does not copy or move data. Instead, it enables Snowflake customers to access shared data at its source. Data sharing is possible because of Snowflake’s decoupled storage and compute layers. This separation enables multiple compute clusters to query the same data set without creating multiple copies. In addition, Snowflake’s automatic scaling eliminates resource competition, so query performance remains high.

Data Sharing can be used for a number of use cases. Organizations can provide customers access to live data as a service. They can also streamline operations by eliminating data silos and intermediary steps typically required for data sharing.

Sharing data requires users to perform a few steps by writing simple SQL statements:

• Create a new data share that specifies the data you want to share.
• Grant permission for the data share to access a Snowflake database and any objects within (e.g., views and schema).
• Add consumer accounts to the share.
• The consumer creates a new database using the share, which references data in its original location

It’s important to recognize that Data Sharing can only occur between Snowflake customers. However, because Snowflake customers pay by storage and compute resources, organizations, at minimum, can view shared data at no cost by registering as a Snowflake user. These users can view data but not query or update it, unless they want to pay computation charges.

Customer Examples. Localytics is a mobile engagement platform company that sells their data to subscribers for target marketing campaigns. Maintaining data pipelines for each subscriber is difficult, but Data Sharing offers a low cost, scalable solution.

Last year, Adobe bought TubeMogul, a platform for planning and buying video advertising. Adobe creates data segments for targeting users that TubeMogul leverages. Data Sharing enables Adobe to provide access to its segment data for TubeMogul without moving data and creating pipelines.